In this portfolio the following playlists from Spotify are used: ‘All Out 80s’, ‘All Out 90s’, ‘All Out 2000s’ and ‘All Out 2010s’. These playlists were made by Spotify and together they contain the 600 biggest songs from 1980 until 2020. It is very interesting to see how taste in (and production of) music can change in these years. Therefore, the four playlists will be compared to each other to find the differences and similarities in the biggest songs over time. I expect the genres of the songs to mostly be pop, r&b/hip-hop and rock, since these are on average the most popular genres. However, the popularity of each genre might change over the years. It might also be interesting to see if many artists ended up in multiple playlists and how several track-level features of hit songs have changed over the years, such as loudness, popularity, danceability or energy.
The playlists were made by Spotify themselves, meaning the playlists are probably quite representative of the years that Spotify was in use (since 2008). But it might not be as representative for the earlier years, since they might have just included songs that were ‘timeless’, meaning they were most listened to after 2008 and not in their own time. A few songs from the corpus stand out. For example, ‘Beggin’’ from Måneskin, since this song is actually a cover of a song from 1967 that was never popular then. Another song that stands out is ‘Stan’ by Eminem since it is a very long song (6.5 minutes) compared to the other songs, so it might also be interesting to look at the popularity of songs and their duration. The songs that are most typical for the time periods are probably the top 5 most played songs.
In this graph, it can be seen that the loudness of hits has significantly increased since 1980. This phenomenon is also called the ‘loudness race’. Since loudness can range between -60 and 0 db, it is very interesting to see that none of the hit songs have a loudness below -20. It is also obvious that there is a correlation between loudness and energy of the music. Here, energy ranges from 0.0 to 1.0 and represents a perceptual measure of intensity and activity. This graph shows that, as the loudness increases, so does the energy. Lastly, there is an obvious increase in the use of explicit language (the blue dots) in songs over the years. The increased loudness, energy and use of explicit language might have had an effect on the popularity the tracks in these playlistst.
In this graph, we can see the distribution of the popularity of tracks in the four playlists. We can see that the tracks from the 2010 playlist are the most popular. The playlist with 2000s songs includes on average less popular tracks than the 2010 playlist and the 90s and 80s playlists contain even less popular songs. It is very interesting to see that there is an increase in popularity of the playlists over the years and that the ‘All Out 2010s’ playlist contains the most popular songs. This might be an effect of the increased loudness, energy, use of explicit language or other changes that happened in music over the years. However, this might also (partially) be caused by the fact that Spotify already existed during the release of these tracks and not during the release of the tracks from 1980-2008.
This chromagram shows the chroma features of Wicked Game by Chris Isaak. Wicked Game is known as a soft rock song and is performed in a sorrowfully conflicted, brooding tone. This song is the least loud song in the corpus (-18.090) and also has low energy (0.296).
As can be seen in the chromagram, the song is mostly in B key and E key. It does not change pitch a lot, except for a few moments when it changes to A, C#|Db or F#|Ab. The fact that the pitch is mostly E and B might attribute to the low energy in the song.
This cepstrogram shows…